Made spaces

README.md

@@ -13,89 +13,85 @@ It’s Machine Learning time with MC-ML!
 
 **VERSE:**
 
-First you got a dataset 
+First you got a dataset  
 and look what you get  
-You gotta a buncha rows, columns, 
-and some headers I bet
+You gotta a buncha rows, columns,  
+and some headers I bet  
 
 
-Before we go through it 
+Before we go through it  
 and we build a machine  
-We gotta make sure your data’s 
+We gotta make sure your data’s  
 sparkling clean  
 
 
-Randomize the rows before you 
+Randomize the rows before you  
 do the next step  
-Split your cleaned up data 
+Split your cleaned up data  
 into training and test
 
 
-Now scale that data, 
+Now scale that data,  
 if you know what I mean  
-Gonna normalize it 
+Gonna normalize it  
 for an unbiased machine
 
 
-Now we choose an algorithm 
+Now we choose an algorithm  
 to predict or classify  
-Both use regression, f(x) = y
+Both use regression, f(x) = y  
 
 
-If you classify, 
+If you classify,  
 make sure your classes are equal  
 If one is bigger than the other then your accuracy is worthless  
-That doesn’t rhyme, but it’s still very important
+That doesn’t rhyme, but it’s still very important  
 
 
-Regularization 
+Regularization  
 is the dopest shit  
-It limits overtraining 
-so we don’t overfit
+It limits overtraining  
+so we don’t overfit  
 
 
-Now score that model, 
+Now score that model,  
 gotta see if it works  
-Count the true and false positives 
-to see what it’s worth
+Count the true and false positives  
+to see what it’s worth  
 
 
-You slide a loss function 
+You slide a loss function  
 then it changes them all  
-Cause your TP rises 
-but your TN falls
+Cause your TP rises  
+but your TN falls  
 
 
-Cross-validate, 
-don’t be a sucka  
-Same set, different tests, 
-10 times muthaf***a
+Cross-validate,  
+don’t be a sucka   
+Same set, different tests,  
+10 times muthaf***a  
 
 -10-
 
-Now excuse me
-If I'm gettin too deep
-We got a neural network here
-With parameter sweeps
+Now excuse me  
+If I'm gettin too deep  
+We got a neural network here  
+With parameter sweeps  
 
 
-We can approximate any 
-continuous function
-Back propagate
-Change the weights for production
+We can approximate any  
+continuous function  
+Back propagate  
+Change the weights for production  
 
 
-Yann Lecun
-And Ian Goodfellow
-They call me big boss
-Cause my hacks so ghetto
+Yann Lecun  
+And Ian Goodfellow  
+They call me big boss  
+Cause my hacks so ghetto  
 
 
-All the ladies love my
-Google Colabs
-
 BREAKDOWN
 
-Yeah, I work for a startup mutha**a!
-They got a fridge full of cheese motha**a!
-Everbody love me when I roll in my Prius (it's so economical)
+Yeah, I work for a startup mutha**a!  
+They got a fridge full of cheese motha**a!